Prenatal prediction of neonatal outcome in the extremely low-birth-weight infant

OBJECTIVE: Our goal was to identify prenatally available parameters that correlate with neonatal outcome and could be used for predicting such outcome in the extremely low-birth-weight pregnancy. STUDY DESIGN: From 1990 through 1995, obstetric and neonatal data of live-born nonanomalous singleton infants with birth weights between 400 and 1000 gm were reviewed. Only cases in which ultrasonographic biometry, including biparietal diameter, abdominal circumference, and femur length, was performed less than or equal to 3 days before delivery were included. Overall survival (defined as alive al discharge) and survival without specific severe neonatal morbidities (namely, retinopathy of prematurity [stage 3 or 4], intraventricular hemorrhage [grade 3 or 4], periventricular leukomalacia, chronic lung disease, and deafness) were ascertained. The best combination of prenatal parameters for the prediction of overall survival and survival without severe morbidity was determined by backward stepwise logistic regression analyses. RESULTS: The most significant prenatal predictors of overall survival were the obstetric estimate of gestational age and the abdominal circumference (chi(2) = 11.8036, p = 0.0006 and chi(2) = 8.1862, p < 0.005, respectively). Survival without severe morbidity was also predicted by the same combination of parameters (chi(2) = 21.9079, p = 0.0001 and chi(2) = 6.538, p = 0.01, respectively). The estimated fetal weight was not a significant independent predictor of either category of outcome (chi(2) = 0.1249, p = 0.72 and chi(2) = 0.0361, p = 0.85, respectively). On the basis of the regression formulas, curves displaying the probabilities of overall survival and survival without severe morbidity with any combination of gestational age and abdominal circumference were developed. CONCLUSION: The combination of gestational age and the abdominal circumference measurements appears to be superior to any combination that included estimated fetal weight data for predicting neonatal outcome in the neonates weighing less than or equal to 1000 gm. We developed a mechanism for predicting neonatal outcome in this weight category on the basis of prenatally available parameters. This information could prove useful for both parental counseling and obstetric decision making.